California Governor Newsom signs executive order pushing AI 'kill switch' requirements and independent lab oversight

California Governor Gavin Newsom signed an executive order convening an expert panel that must deliver AI safety recommendations within two months, including a routinely verified 'kill switch' for frontier models, onsite independent verifiers, audited…

California Governor Gavin Newsom signed an executive order (reported 2026-09-18; The Verge describes it as a Friday order) convening an expert group that must deliver frontier AI safety recommendations within two months. Measures under consideration include onsite independent verification groups — which The Decoder describes as independent auditors embedded directly inside AI labs — independently audited transparency reports, a routinely verified 'kill switch' for frontier models, and mandatory reporting of loss-of-control incidents, with both outlets citing the OpenAI attack on Hugging Face as an example incident. The order also accelerates two recently signed California laws establishing an independent AI verifier framework and a state registry of AI auditors. Newsom urged Congress to adopt California's framework as a national baseline or regulatory floor, noting that no federal law currently requires AI companies to report dangerous incidents.

• Governor Gavin Newsom signed an executive order reported on 2026-09-18 (described by The Verge as a Friday order).
• An expert group/panel must deliver AI safety recommendations within two months (stated in both reports).
• Proposed measures include onsite independent verification groups (The Verge) / independent auditors embedded directly in AI labs (The Decoder).
• The Verge says proposed measures also include independently audited transparency reports and a routinely verified 'kill switch' for frontier models.
• Companies would be required to report loss-of-control incidents; both reports cite the OpenAI attack on Hugging Face as an example.
• The order accelerates two recently signed California laws: an independent AI verifier framework and a state registry of AI auditors (The Verge).
• No federal law currently requires AI companies to report dangerous incidents (The Decoder).
• Newsom urged Congress — and, per The Verge, also Trump — to adopt California's framework as a regulatory floor/national baseline.